Monday, August 29, 2011

Report the name of a JavaScript function

I do a LOT of JS development lately: straight up HTML5, PhoneGap for Android and iPhone and Node.JS on the server side. One command line I use in all environments is console.log.
console.log allows you to print out a string, a variable and even a complex object (just use JSON.stringify on it first) to the console view of:
  • Developer Tools in IE8 and higher (hit F12)
  • Developer Tools in Chrome (different keys for different OSs – nice, Google)
  • Firebug in Firefox (install as an extension)
Since advanced JS uses a lot of event calls, anonymous functions and multiple files, it’s hard at times to figure out where things go wrong. So I’ve gotten into the habit of adding an entrance printout at the beginning of each function. But it has gotten cumbersome and annoying after a while, so I searched for a more elegant solution. I found this StackOverflow question, and based on the replies, constructed this code:

var fname = /^function\s+([^(]+)/.exec(arguments.callee.toString())[1];
console.log(fname);


Now, this looks like a cumbersome code to paste into each and every function, so I wanted to write a general funnction I can call anywhere to print the current function name. The problem is, if you put this code into a function, say reportFunctionName() and call it, you get "getFunctionName" printed out. The answer was in the same SO post:

function reportName() {
  var fname = /^function\s+([^(]+)/.exec(reportName.caller.toString())[1];
  console.log(fname);
}


Note the use of caller instead of callee. Also note both methods would not work well for anonymous functions (duh - no name -> nothing to print :)

Finally, I started using Github's Gist feature to collect all my code snippets, so you can find both code pieces mentioned here and here respectively.

Monday, August 15, 2011

Ultraviolet Pig

If you got to this post hoping to read about the “Angry Birds” sequel “Ultraviolent Pigs” – you’re in the wrong place - please go to the Rovio siteSmile.
This is just another crazy electronics project I set myself to.

It all started while I was zombie-surfing the DealExtreme site. If you’re not familiar with the concept, it means browsing after 1am, no good reason other then to eke a couple of more minutes of browsing before the brain shuts down completely. Instead of muttering “brrraaaaains!” you mutter “beeedd”. Some of my best (and worst) ideas hit me during zombie surfing. But I digress.

So, as I said, while zs-ing DealExtreme, I came a cross a too-stupid-not-to-buy item: a pig LED flashlight for $0.78. And free shipping to boot! What’s the catch here, I wondered? And what the hell will I do with this useless, clearly non-kosher flashlight?


The Ultraviolet Pig
Here piggy piggy!

So I browsed further, and for $2.93 I got a bag full of ultraviolet LEDs.

20 UV LEDs
Not the best UV LEDs out there - but the cheapest
After 20-25 days (hey, you 'pay' for free shipping, you get it delivered by a lame horse that has to swim across the Pacific blindfolded), I finally got an envelope containing the items.

It took less than 5 minutes to take the pig apart (please don’t ever associate this sentence with me Smile) and replace the 2 simple LEDs with the UV ones. And that’s when the fun starts.


Other than enabling me to see hidden marks on a $20 bills and a driver’s license, the most useful thing you can do with UV light is protect your stuff.
Do you have a favorite chair that you think gets replaced with a crappy one when you leave your desk? Did someone swipe your beloved Swingline stapler and you can’t prove it? Just mark your stuff with UV ink – completely invisible to the naked eye – but not to good ol’ Mr. UVP. Observe:

Now you don't

This is a simple metal mint box, easily swiped and easy to deny. But when you turn UVP on it, the truth comes to light, in 3 colors (you can also see the UV pens in the photo):

Now you see it


So, yes, not really practical unless you can retrieve the stolen goods, but quite good at identifying your stuff, or secretly communicating under the nose of pig-less people (think about passing notes in a boring meeting).

The stats:
  • Usefulness: 3/5 stars – not that amazing, and not too techy. Not a single line of code needed
  • Fun factor: 5/5. Oh yeah!
  • How easy it is to build: very. Just get your pig, LEDs (you need at least 2) and follow your instincts. Or watch this youtube video where someone demonstrates how to take it apart
  • Total project cost: $3.71 – and I still have about 18 LEDs waiting for the next 9 pigs

Warning: DO NOT STARE DIRECTLY INTO UV LIGHT.
Your eyes cannot see it well, hence your pupils will not shrink properly to protect them. Few minutes are OK, more is not recommended.

Thursday, July 7, 2011

The 5-minutes Twilio App

This week, I interviewed for a position at Twilio. Twilio sells an API that enables your web applications and mobile app to use voice and SMS messaging capabilities. You are charged a small fee per incoming/outgoing call/message, and you get full backend support, including IVR, call statistics and the ability to reply automatically based on content and context.

The recruiter suggested I build something using the Twilio API. Since I have so many APIs running in my head these days, I was a bit hesitant at first to learn another one, especially if I didn’t see an immediate use for it. I was further less impressed by the use of XML (over JSON) in the API. But I gave it a shot, and within 5 minutes of starting I had my one (and only) Twilio API written: the SMS Calculator (no tm or c need apply). One minute after that and I had it tested. 5 minutes after twitting about it, someone asked me if he can demo my app in a code meetup. Once I said ‘yes’, my app officially became open sourced.

So, without further ado, here’s the code to my app, and the steps you need to take to create your 5-minutes Twilio app in 5 easy steps:

1. Go to twilio.com and create a free account. All you need is an email and a password.
You get free $30 on your account to use for testing. Good for 3000 SMS messages.

2. Paste the following PHP code into a text file and name it calc.php

<?php
 $body = $_REQUEST['Body'];
 $expression = explode(" ", $body, 3);
 $a = intval($expression[0]);
 $b = intval($expression[2]);
 $result = 0;
 switch($expression[1]) {
  case "+":
   $result = $a + $b;
   break;
  case "-":
   $result = $a - $b;
   break;
  case "*":
   $result = $a * $b;
   break;
  case "/":
   if($b != 0)
    $result = $a / $b;
   else
    $result = "Division by 0";
   break;
  default:
   $result = "ERROR";
   break;
 }
 $reply = "You asked: $body. The reply is: $result";
 header("content-type: text/xml");
 echo "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n";
?>
<Response>
 <Sms><?php echo $reply ?></Sms>
</Response>

Code explanation:
This is a rudimentary, "stoopid", expression calculator. It expects an operand-space-operator-space-operand syntax (i.e. 2 + 2), does not use any regular expressions and does not check for correctness. The only error it checks for explicitly is division by 0 (and that was added only because I had points deducted for a similar exercise years ago by an evil TA).
  • We get the SMS body in the $_REQUEST[‘Body’] HTTP variable.
  • We proceed to split that string, using space as a delimiter. I limit the split to 3 strings.
  • We parse the 2 expected integers
  • We attempt to parse the operand (only the main 4 are supported)
  • We calculate the result and compose the reply. I could probably use more string tricks here, but what’s the use?
  • Finally, we compose the SMS XML that Twilio expects and output it.

3. Put the page on a web server that supports PHP. There are plenty of free hosting choices if you don’t have one ready (off the top of my head, try Zymic)

4. In your Twilio account page, scroll down to the section “Developer Tools”. Input the full path to the calc.php page in the “SMS URL” field, and notice the phone number and PIN you were allocated.

Twilio

At this point you’re actually done. Step 5 is the testing.

5. Using your phone, send an SMS message to the phone number. Start by putting in your PIN number (only required with test accounts) and then add an expression. It should look like:
xxxx-xxxx
25 / 5
You should receive an immediate reply, echoing the given expression and the result. For the full free experience, test using your Google Voice account. Google provides free SMS texts in the US and Canada, and the Twilio number is a US one.

I will certainly keep Twilio in mind for future uses in my apps, but as for now, I'm just happy that it took me longer to write this post than to write the code.

My next step: slapping an Eliza or an 8-ball app at the end of this :)

Update:

Andrew Watson from Twilio asked to borrow my sample for his presentation. Here's a link to the Google Doc Andrew has published. Thanks Andrew!

Oh, and it seems like I won't be working for Twilio after all, so I'm free to play with other new toys :)

Thursday, May 19, 2011

Display JSON results

JSON (JavaScript Object Notation) is a great straight forward way to transfer information from one application/server to another. It's string-based, array-oriented, and does not carry with it the redundancy of XML or (shudder) SOAP. And the best feature in a web application: it's immediately transferable to a JavaScript object, without the need for (an explicit) parsing.

Add to the mix jQuery - a JavaScript framework that makes using JavaScript and writing dynamic web pages a snap. Once I picked it up, I couldn't go back. jQuery is now embraced by many - including Microsoft and Google, who include it in their default web templates, and provide hosted CDN versions of its latest versions.

jQuery shines when it comes to AJAX transactions - you can get a JSON result from a web service using one instruction! Going over the reply is also a snatch. I found myself many times just needing to dump the results on the page - either for debugging purposes, or so I could copy/paste them somewhere.

Here is a short code snippet that I use to get a JSON result from a local web service (a remote request requires JSONP - future post), and dumps the results in a form of a list. It assumes a single-level array reply and that you have a <DIV> element on the page called "result":


        


Explanation:
  1. Line 1 includes version 1.5.2 of jQuery from the Microsoft CDN
  2. Line 4 will make sure the function will run when the document has been fully loaded and drawn
  3. Line 5 calls the JSON service and sends the results to a function
  4. Line 8 iterates over the result object and creates an array of <LI> elements
  5. Line 14 creates a new <UL> element, applies a CSS style to it, attaches the <LI> elements and places it inside the result <DIV>
That simple!

Monday, May 2, 2011

Netduino Project 2: Morsey

As discussed in my last post, I've created a small test app called Morsey to practice my LED lighting skills on the Netduino.

The functionality is simple: the application takes a constant English string and turns it into a string of dashes and dots, signaled by the LED. The signaling is started/stopped by clicking the on board button.

The code:
There are 3 code files:
  1. The main program file, that runs the main thread and the Morse thread, as well as the button event.
    static void button_OnInterrupt(uint data1, uint data2, DateTime time)
    {
     bool isButtonDown = (data2 != 0);
     if (isButtonDown)
     {
      displayMessage = !displayMessage;
     }
     
     if (displayMessage)
     {
      if(morseThread.ThreadState == ThreadState.Unstarted)
       morseThread.Start();
      else
       morseThread.Resume();
     }
     else 
     {
      morseThread.Suspend();
      led.Off(); 
     }
    }
    
  2. The MorseCharacters class. I copied the 2 Morse arrays from Scott Hanselman's great Netduino post (sorry Scott, just lazy and it's faster to C&P than to type all those dashes and dots :)).
  3. The LED class - the same one I used in the TrafficLight project, with the addition of a single function:
    public void WriteMorse(String message) 
    {
     MorseCharacters converter = new MorseCharacters();
     String morseMessage = converter.ConvertMessageToMorse(message);
     if (morseMessage != String.Empty)
     {
      foreach (char c in morseMessage)
      {
       switch (c)
       {
        case '.':
         Dot();
         break;
        case '-':
         Dash();
         break;
        case ' ':
         //delay for twice the character gap between words
         Thread.Sleep(delay * 2); 
         break;
        default:    //handle errors here
         break;
       }
      }
     }
    }
    

To get the full code, clone the git at https://github.com/TravelingTechGuy/Morsey

Sunday, May 1, 2011

1st Netduino Project: Traffic Light

While I'm dabbling in HTML5, iPhone and Android projects at this particular point in time, I chose to open this blog with a project I'm passionate about. I've been following the Arduino revolution for years now, but always from a far. Although I have some electronics background in my ancient past (try: middle school electronics kits), I see myself a s a software professional. Having had the distinct "pleasure" of working in assembler, I decided that's as close I'd ever get to speaking to real hardware.

And then I found out about the Microsoft .Net micro framework. It seems I can take my .Net skills and transfer them to the hardware world! I can make LED lights turn on and off! I can react to button clicks rather than keyboard and mouse events! But on a more serious note, I can implement a standalone web server that allows me to perform operations remotely. I can read/write to an SD card. I can control a vast collection of electronic components and analyze information from a myriad of sensors. All this with a very small upfront investment. I'm in.

What do you need:
  1. An arduino board that supports the .Net micro framework. There are several out there, but I chose the Netduino Plus. At $60 it provides an Ethernet port and a microSD card reader on board, saving you the need to wire those later. I got it off Amazon, along with the breadboard (see 3).
  2. Free software: SDK and drivers (most software associated with Arduino is open source, even (shock!) the .Net micro framework from Microsoft. This page has a full list of what you'll need, including a good installation guide that will take you through installing, writing your first program and debugging it. 
  3. To develop software you use Visual Studio (you can use the free Visual C# Express version). 
  4. Some cheap components. I got a mini breadboard for $3. A breadboard allows you to play with circuits and components without the need to solder anything. There are no mistakes and most components are tolerant with the involved voltages. Heck, I found out that one of the LEDs I got was dual-colored - you get a different color if you plug it in the other way. That was a nice surprise.
    I also got a couple of wires and color LEDs from my local FRY's (at $0.49 each - you can get them at $5/40 - I just didn't need that many).
  5. Plenty of internet reading: instructions, projects, suggestions. I enrolled in the Netduino forums and have been so far assisted with some issues. Every one there is enthusiastic and helpful. All are hobbyists - none are trolls.
  6. A free weekend.
Later on you may want to buy more components, like shields (Arduino add-on boards) sensors (light, touch, accelerometers), wireless network adapter (like XBee), LCD screens, joysticks - whatever.
My recommendation is: even though the components are relatively cheap (with the exception of XBee and a GPS shield, they are almost always below $20, most are even cheaper), avoid buying them until you feel sure you've grasped the basics and are ready to proceed.

Programming:
The "Hello, World" project of  the Arduino world is called "Blinky". You get to blink the on-board LED, and later read the on-board button to control the LED.

Visual Studio automates the deployment of  allows you to do a step-by-step debugging on the device. This is an amazing way to develop for hardware. Of course, you get the assortment of tools you're already used to as a .Net developer.

I asked myself what can I do with an LED, and came out with the done-to-death idea of a Morse code Signal Emitter (I called it "Morsey"). You feed it an English string and it yields a string of dashes and dots using the LED. The next step was to use the button to start/stop the signaling. And then came the notion of using a separate thread to be able to pause and resume the signaling - in short, I spent time honing my skills, such as they are, preparing for more complex projects.

Traffic Light project:
But before I was ready to move on to my next project, which will involve an LCD display ($10 on Amazon), I decided to test my breadboard skills and developed this small project called "Traffic Light".
Made of one tiny breadboard, 3 colored LEDs, 7 wires and about 40 lines of code, it simulates a traffic light's color change, using a separate thread, and the LED class I carried over from "Morsey".

You can see the end result in this video and I should really start to remember not to talk while zooming with this camera - it records the zoom noise over your voice):



I found a nice (free) application called Fritzing that allowed me to design my circuit design and share it with other hobbyists. Here's what the above circuit looks as a Fritzing design:
And Fritzing automatically turns it into a schema you can use to share with electronics professionals, using standard conventions. Just in case you'd like to mass-produce your project one day:

If you want to get the full code, including these images and the Fritzing file, just git it here: https://github.com/TravelingTechGuy/TrafficLight.

And yes, I wrote "Git", not "get". If you don't know what Git is, or how to use it, the page also provides a zip download, but you better get used to it, as I'm going to use github to share all my code on this blog (and thanks to Jason for introducing me to git in the first place).

Well, this is done, on to the next project.